Adolescent Striving and Methylation Aging
Lauren Gaydosh1, Colter Mitchell2, Lisa Schneper3,Daniel Notterman3, and Sara McLanahan3
1University of North Carolina at Chapel Hill
2University of Michigan
3Princeton University
April 27, 2018
Finding
• Adolescents from highly disadvantagedbackgrounds who strive to succeed in schoolexperience more rapid epigenetic aging acrosschildhood and adolescence
• This pattern is observed among black andHispanic adolescents but not among whites
Motivation
• Evidence that higher levels of self-control, striving, andsocioeconomic success among disadvantaged blacks andHispanics are associated with worse physical health
I John-Henryism (James 1983)
I Skin-deep resilience (Brody et al. 2013; Miller et al. 2015)
I Unequal benefits (Gaydosh et al. 2018)
Research questions
1 What is the relationship between early life disadvantage,childhood striving, and the pace of epigenetic aging?
2 Does this relationship vary by race/ethnicity?
Data
• Probability sample of 4,898 births in 75 hospitals across 20US cities
• Oversample of non-marital births
• Baseline survey at birth
• Follow-up at years 1, 3, 5, 9, and 15
Data - DNA methylation
• Saliva collected at years 9 and 15 using Oragene DNAself-collection kit
• Preliminary sample of n ∼ 700 respondents using IlluminaInfinium HumanMethylation 450K BeadChip
• Stay tuned: will add more samples using EPIC chip, n > 2,000
Measures - DNA methylation (DNAm) age
• Horvath method for calculating epigenetic age:I Analyzed 21k CpG sites across 51 different tissue and cell
types from over 7k samplesI Regressed chronological age on the CpG sites using a penalized
regression modelI Identified 353 CpG sites as “clock” CpG sitesI Robust across tissue types
• Construct DNAm age using the weighted average ofmethylation levels at these CpGs
Measures - DNAm age and chronological age
• Modest correlation between DNAm age and chronological age
• Little variation in chronological age in cross-section
• Exploit greater variation over time by looking at pace ofepigenetic aging:
Pace = (DNAm age 15−DNAm age 9)/(Date15−Date9)
Measures - pace of methylation aging
Measures - pace of methylation aging
Measures - socioeconomic disadvantage at birth
• Baseline interview
• 4 indicators at family level: household income below povertyline, receipt of public assistance, mother’s education < HS,single parent household
• 7 indicators at neighborhood level: welfare receipt,unemployment, poverty, educational attainment, femaleheaded, segregation, vacancies
• Summed across indicators
• Mean = 4.5, SD = 3.0, α = 0.79
Measures - striving across childhood andadolescence
• Years 9 and 15 child surveys
• 12 measures of educational aspirations, school engagement,perseverance, and optimism
• Highest response to each question considered “striving”,summed across measures, averaged between years
• Mean = 5.2, SD = 2.2, α = 0.89
Methods
• Linear regression model predicting annual pace of epigeneticaging
• Interaction between disadvantage and striving
• Models control for sex, race/ethnicity, and pubertaldevelopment
Results
Results
Conclusion
• Pace of DNAm aging is accelerated for disadvantagedadolescents who strive academically
• Suggestive evidence that this may be restricted to black andHispanic adolescents
• Physical health risk associated with striving amongdisadvantaged minorities may appear earlier in the life coursethan previously documented
Acknowledgments
Fragile Families is funded by:
• Eunice Kennedy Shriver National Institute of Child Health andHuman Development (NICHD): R01 HD076592,R01HD36916, R01HD39135, and R01HD40421
• National Institute of Mental Health (NIMH): R01MH103761
• Consortium of private foundations
Thank you.Questions: [email protected]